Symbiotic nutrient cycling enables the long-term survival of Aiptasia in the absence of heterotrophic food sources

Phototrophic Cnidaria are mixotrophic organisms that can complement their heterotrophic diet with nutrients assimilated by their algal endosymbionts. Metabolic models suggest that the translocation of photosynthates and their derivatives from the algae may be sufficient to cover the metabolic energy demands of the host. However, the importance of heterotrophy to the nutritional budget of these holobionts remains unclear. Here, we report on the long-term survival of the photosymbiotic anemone Aiptasia in the absence of heterotrophic food sources. Following one year of heterotrophic starvation, these anemones remained fully viable but showed an 85 % reduction in biomass compared to their regularly fed counterparts. This shrinking was accompanied by a reduction in host protein content and algal density, indicative of severe nitrogen limitation. Nonetheless, isotopic labeling experiments combined with NanoSIMS imaging revealed that the contribution of algal-derived nutrients to the host metabolism remained unaffected due to an increase in algal photosynthesis and more efficient carbon translocation. Taken together, our results suggest that, on a one- year timescale, heterotrophic feeding is not essential to fulfilling the energy requirements of the holobiont. But, while symbiotic nutrient cycling effectively retains carbon in the holobiont over long time scales, our data suggest that heterotrophic feeding is a critical source of nitrogen required for holobiont growth under oligotrophic conditions.

Phototrophic Cnidaria are mixotrophic organisms that can complement their heterotrophic diet with nutrients 28 assimilated by their algal endosymbionts. Metabolic models suggest that the translocation of photosynthates 29 and their derivatives from the algae may be sufficient to cover the metabolic energy demands of the host. 30 However, the importance of heterotrophy to the nutritional budget of these holobionts remains unclear. Here, 31 we report on the long-term survival of the photosymbiotic anemone Aiptasia in the absence of heterotrophic 32 food sources. Following one year of heterotrophic starvation, these anemones remained fully viable but 33 showed an 85 % reduction in biomass compared to their regularly fed counterparts. This shrinking was 34 accompanied by a reduction in host protein content and algal density, indicative of severe nitrogen limitation. 35 Nonetheless, isotopic labeling experiments combined with NanoSIMS imaging revealed that the contribution 36 of algal-derived nutrients to the host metabolism remained unaffected due to an increase in algal 37 photosynthesis and more efficient carbon translocation. Taken together, our results suggest that, on a one-38 year timescale, heterotrophic feeding is not essential to fulfilling the energy requirements of the holobiont. 39 But, while symbiotic nutrient cycling effectively retains carbon in the holobiont over long time scales, our data 40 suggest that heterotrophic feeding is a critical source of nitrogen required for holobiont growth under 41 oligotrophic conditions. 42

Introduction 43
Photosymbiotic Cnidaria, such as corals and anemones, dominate shallow hard-bottom substrates in the 44 oligotrophic tropical ocean (Pandolfi 2002). The key to their evolutionary and ecological success under these 45 conditions lies in their association with endosymbiotic algae of the family Symbiodiniaceae (Stanley 2006;46 Stanley and van de Schootbrugge 2009). Efficient nutrient exchange in these symbioses couples the 47 heterotrophic metabolism of the host with the autotrophic metabolism of their algal symbionts ( as our understanding of potential prey dynamics (e.g., zooplankton abundance) and cnidarian grazing on coral 70 reefs remains limited (Lowe and Falter 2015), the importance of heterotrophic nutrients for sustaining the 71 stable cnidarian-algal symbiosis is less clear. 72 Here, we performed a starvation experiment using the photosymbiotic sea anemone Aiptasia to study the role 73 of heterotrophic nutrient acquisition in symbiosis. For this, we reared Aiptasia for one year in the absence of 74 any heterotrophic nutrient sources. This permitted us to examine the effects of heterotrophic starvation on 75 the symbiosis in light of the underlying carbon and nitrogen cycling and explore the limits of autotrophic 76 nutrient acquisition in the photosymbiotic Cnidaria. 77

Animal husbandry & experimental design 79
The experiments and measurements were performed on the photosymbiotic cnidarian model organism 80 Aiptasia, i.e., Exaiptasia diaphana (Puntin et al. 2022 salina nauplii (Sanders GSLA, USA) followed by a complete water exchange and removal of biofilms. 88 For the experiment, all animals were reared under the same conditions as outlined above for one year. 89 However, while half of the animals (two culture containers with five animals each) were fed weekly with 90 Artemia nauplii (regularly fed control), the other half (two culture containers with five animals each) was 91 reared in the absence of any food sources (heterotrophically starved). Apart from this, all culturing parameters 92 were kept identical, including the weekly cleaning and water exchange. Pedal lacerates of Aiptasia were 93 continuously removed to avoid experimental biases due to differences in asexual reproduction and resulting 94 densities of animal populations. 95

Phenotypic characterization 96
Following the one-year experiment, treatment responses were recorded. First, photos of representative 97 phenotypes for each of the treatments were taken with an OM-1 camera and a 60 mm f2.8 macro-objective 98 (OM System, Japan) using identical illumination and exposure settings. Then, three animals were collected 99 from each treatment group, transferred to a pre-weighed 1.5 mL Eppendorf tube, and homogenized in 500 µL 100 Milli-Q water using a PT1200E immersion dispenser (Kinematica, Switzerland). 101 Host and algal symbiont fractions were immediately separated by centrifugation (3000 g, 3 min, sufficient to 102 remove > 95 % of algal symbionts from the supernatant) and the host supernatant was transferred into a new 103 pre-weighed 1.5 mL tube, flash-frozen in liquid nitrogen, and stored at -20°C for later analysis. The algal 104 symbiont pellet was resuspended in 500 µL Milli-Q water and rinsed by one additional centrifugation and 105 resuspension step. Algal symbiont concentrations were quantified in three technical replicates per sample 106 based on cell shape and chlorophyll autofluorescence using a CellDrop cell counter (DeNovix, USA). The 107 protein content in the defrosted host supernatant was quantified in three technical replicates using the Pierce 108 Rapid Gold BCA Protein Assay Kit (Thermo Scientific, USA) according to the manufacturer's instructions. Algal 109 concentrations and host protein content were extrapolated to the initial sample volume and normalized to 110 holobiont biomass. Holobiont biomass was approximated as dry weight. For this, host and symbiont fractions 111 were dried at 45°C until the weight was stable and the initial weight of empty tubes was subtracted from the 112 final weight. The weight of host and symbiont fractions was corrected for aliquots taken for sample 113 measurements to approximate the dry weight of the holobiont as a whole, i.e., host + symbiont fraction. 114

Isotope labeling & sample processing 115
To study treatment effects on symbiotic interactions, we quantified inorganic carbon and nitrogen assimilation 116 and translocation in the symbiosis. For this, three animals from each treatment were transferred to 50 mL 117 glass vials. For isotopic labeling, vials were filled with minimal artificial seawater medium (35 PSU, pH 8.1, 118 355.6 mM NaCl, 46.2 mM MgCl2, 10.8 mM Na2SO4, 9.0 mM CaCl2, 7.9 mM, KCl; (Harrison et al. 1980)) 119 containing 2.5 mM NaH 13 CO3 and 10 µM 15 NH4Cl. In addition, one additional animal per treatment was 120 transferred into a vial filled with minimal artificial seawater medium without heavy isotope tracers to serve as 121 unlabeled controls for NanoSIMS measurements. Animals were incubated for 6 h in the light at their regular 122 culture conditions before being transferred to a fixative solution (2.5 % glutaraldehyde and 1 % 123 paraformaldehyde in 0.1 M Sorensen's phosphate buffer). Samples were fixed for 1 h at room temperature 124 followed by 24 h at 4°C before being stored in a preservative solution (1 % paraformaldehyde in 0.1 M 125 Sorensen's phosphate buffer) at 4 °C until further processing. Within four days of fixation, samples were 126 dissected and individual tentacles were processed for resin embedding. Following secondary fixation in 1 % 127 OsO4 for 1 h, samples were rinsed (3 x MiliQ for 10 min) and dehydrated in a series of increasing ethanol 128 concentrations (30 % for 10 min, 50 % for 10 min, 2 x 70 % for 10 min, 3 x 90 % for 10 min, and 3 x 100 % for 129 10 min) before being transferred to acetone (100 % for 10 min and transferred onto glow-discharged silicon wafers. 134

Electron microscopy and NanoSIMS imaging 135
For scanning electron microscopy (SEM), sections were stained with uranyl acetate (1% for 10 min), followed 136 by Reynolds lead citrate solution (10 min). Images of tentacle of sections were taken on a GeminiSEM 500 field 137 emission scanning electron microscope (Zeiss) at 3 kV with an aperture size of 30 µm, and a working distance 138 of 2.8 mm, using the energy selective backscatter detector (EsB; ZEISS) with the filter-grid set at 121 V. For 139 NanoSIMS imaging, sections on wafers were sputter-coated with a 12 nm gold layer using an EM SCD050 (Leica 140 Microsystems) and the samples were analyzed with a NanoSIMS 50L instrument (Hoppe et al. 2013). To 141 remove the metal coating, target sample areas were pre-sputtered for 5 min with a primary beam of ca. 6 pA. were analyzed across two animal replicates. For isotopically labeled Aiptasia, 120 host and 120 algal symbiont 157 ROIs were analyzed across three animal replicates per treatment. Based on this, the 13 C and 15 N assimilation 158 in individual ROIs was expressed as atom % excess (in comparison to corresponding ROIs of unlabeled control 159 Aiptasia). Due to the clonal nature of Aiptasia and the identical environmental conditions of animals within 160 the same treatment, individual ROIs were considered independent measurements across animal replicates for 161 the purpose of this study. 162

Statistical analyses 163
Treatment effects on phenotypic responses, i.e., biomass, host protein content, and symbiont density, were 164 analyzed using two-sided unpaired Student's t-tests. Isotope ratios from NanoSIMS analysis were log-165 transformed to meet model assumptions and analyzed with linear models (LM) using the respective symbiotic 166 partner (host/symbiont) and treatment (fed/starved) as explanatory variables. To test individual differences 167 between groups LMs were followed up with a Tukey HSD post hoc comparison. 168

Holobiont biomass loss in the absence of heterotrophic nutrients 170
After one year of husbandry in the absence of heterotrophic food sources, no mortality was observed and 171 Aiptasia remained viable but had ceased any detectable asexual propagation via pedal lacerates. Starved 172 animals showed pronounced phenotypic differences compared to their regularly fed counterparts. 173 Specifically, starvation resulted in a reduction in body size, a paler appearance, and a loss of 85 % of their dry 174 weight (Fig. 1A,B; Student's t-test, t = 4.71, p = 0.042). This decline in holobiont biomass was, at least in part, 175 driven by a strong decline in host protein content and algal symbiont density, which both decreased by more 176 than 80 % on average when normalized to holobiont biomass (Fig. 1C

184
Enhanced photosynthetic performance of algal symbionts sustains host metabolism during heterotrophic 185 starvation 186 SEM images of tentacle sections confirmed that Aiptasia from both treatments hosted algal symbionts in their 187 gastrodermal tissue. However, gastrodermal cells from starved Aiptasia appeared to contain a higher density 188 of lipid bodies than those of their regular fed counterparts ( Fig. 2A,B, Fig. S1). Likewise, algal symbionts from 189 starved Aiptasia appeared to have a higher lipid content in their cells than those from regularly fed Aiptasia, 190 albeit this trend was less clear than in the host tissue (Fig. S1). densities during heterotrophic starvation, overall 13 C enrichment remained stable in the gastrodermal tissue 198 of the host (Tukey's HSD, p = 0.951). This was best explained by the enhanced photosynthetic performance of 199 algal symbionts, which exhibited an increase of nearly 50 % in 13 C enrichment in starved animals (Fig. 2E, Fig.  200 S2; Tukey's HSD, p < 0.001). in both epidermal and gastrodermal tissue layers (Fig. 2F,G, Fig. S2; host/symbiont differences: LM, F = 206 3320.37, p < 0.001). Thus, heterotrophic starvation did not alter the ability for ammonium assimilation of 207 either symbiotic partner (Fig. 2H, Fig. S2; Tukey's HSD, p = 0.489 for host gastrodermis, p = 1.000 for algal 208 symbionts).

Discussion 220
The association with autotrophic endosymbiotic algae has enabled heterotrophic Cnidaria to thrive in the 221 oligotrophic tropical ocean (Muscatine and Porter 1977;Stanley 2006). The long-term starvation experiment 222 presented here emphasizes the remarkable trophic plasticity that such a symbiosis confers upon these 223 cnidarian holobionts. Because of the highly efficient symbiotic nutrient exchange and recycling, Aiptasia were 224 able to survive without heterotrophic feeding for at least one year. At the same time, starved animals showed 225 clear signs of nutrient limitation, including reduced biomass, host protein content, and symbiont density, 226 underscoring the long-term importance of heterotrophic feeding for body mass maintenance and growth. 227 Autotrophic nutrient recycling can sustain the cnidarian-algal symbiosis for extended periods of time 228 Recent work suggests that the lack of heterotrophic feeding could shift the cnidarian-algal symbiosis towards 229 parasitic interactions that reduce the capacity of the host to survive starvation (Peng et al. 2020). However, 230 here we show that, even after one year of complete heterotrophic starvation, the translocation of 231 photosynthates by algal symbionts remained sufficient to maintain the basal metabolic requirement of the 232 host. Indeed, patterns of host 13 C enrichment ( Fig. 2A-C) were not affected by heterotrophic starvation 233 indicating that photosynthate availability for the host was not impaired despite an 85 % reduction in algal 234 symbiont biomass in the holobiont (Fig. 1B). Combined with the pronounced increase in the density of lipid 235 bodies in the host tissue of starved Aiptasia (Fig. S2), this implies that carbon translocation by individual algal 236 cells must have significantly increased in response to heterotrophic starvation. Indeed, we observed a 50 % 237 increase in 13 C enrichment and an increased density of lipid bodies among the algal symbionts in starved 238 animals (Fig. 2C), clearly indicating enhanced photosynthetic performance and excess fixed carbon availability 239 required for higher relative translocation rates. Similar, albeit less pronounced, trends were previously 240 reported in a three-month starvation experiment using Aiptasia (Davy and Cook 2001). These authors 241 proposed that the increase in algal photosynthetic performance in starved animals was the result of reduced 242 intra-specific competition for CO2. Indeed, reduced algal symbiont densities likely reduce competition for CO2 243 (Rädecker et al. 2017; Krueger et al. 2020). However, in starved animals, this effect could, in part, be masked 244 by the reduced catabolic CO2 production in the holobiont due to the lack of heterotrophic prey digestion by 245 the host. Our data point to an additional mechanism that could promote enhanced photosynthate release by 246 algal symbionts in the absence of heterotrophy, namely nitrogen starvation. 247

Nitrogen limitation shapes the starvation response of Aiptasia 248
In the stable symbiosis, low nitrogen availability limits the anabolic incorporation of photosynthates in the 249 algal symbiont metabolism (Rädecker et al. 2021;Cui et al. 2022aCui et al. , 2022b. This nitrogen limitation is thus not 250 only crucial in regulating algal growth but also ensures the translocation of excess photosynthates to the host 251 (Muscatine and Porter 1977;Falkowski et al. 1984). The host passively modulates in hospite nitrogen 252 availability for algal symbionts through ammonium assimilation and production of ammonium in its glutamate 253 metabolism (Rahav et al. 1989;Rädecker et al. 2021;Cui et al. 2022a). Here, we found a proportional decline 254 of algal symbiont density and host protein content in heterotrophically starved Aiptasia (Fig. 1C,D) Capone 2008). In this context, our findings illustrate the importance of heterotrophic feeding by the host for 274 the long-term maintenance of the cnidarian-algal symbiosis biomass. While symbiotic nutrient exchange and 275 recycling may be sufficient to cover the carbon and energy demands of the symbiotic partners on a time scale 276 of at least one year, heterotrophic feeding is not only required for long-term survival but also required for 277 propagation and net growth of the holobiont. 278

Conclusion 279
This study has illustrated the substantial trophic plasticity of the cnidarian-algal symbiosis: Aiptasia survived 280 for an entire year in the complete absence of heterotrophic feeding. Our findings reveal that efficient 281 symbiotic nutrient exchange and recycling are sufficient to sustain the basic metabolic requirements of both 282 symbiotic partners over extended periods of time. Yet, under long-term exposure to highly oligotrophic 283 conditions, the assimilation of environmental inorganic nitrogen is not sufficient to support the nutritional 284 requirements of the holobiont, and heterotrophic feeding represents an essential source of nitrogen for 285 holobiont growth.,Mixotrophy thereby provides a nutritional advantage to photosymbiotic cnidarians that in 286 part explains their ability to outcompete other organisms restricted to either autotrophic or heterotrophic 287 nutrient acquisition under oligotrophic conditions. 288

Acknowledgments 289
We are grateful to Dr. Claudia Pogoreutz and Gaëlle Toullec for their help with animal culture maintenance. 290 We thank Jean Daraspe and Dr. Cristina Martin-Olmos for their help with sample processing and SEM imaging 291 as well as Dr. Stéphane Escrig and Florent Plane for their assistance and support with NanoSIMS 292 measurements. We thank the editor/recommender as well as the two reviewers for their constructive 293 feedback that significantly improved the manuscript.